Multi-parametric phospho-flow cytometry: A crucial tool for T lymphocyte signaling studies

Authors


Abstract

Tools such as protein immunoblotting have proven benefits for investigating T lymphocyte signaling but have several drawbacks such as the number of cells required and the difficulty of distinguishing subset-specific differences without expensive and invasive cell sorting. Recent advances in immunology and the identification of T lymphocyte sub-populations making up only a very small fraction of the total population highlight the importance of studying signaling in those small subsets in a feasible, cost-effective, high-throughput manner. To this end, we have developed a simplified protocol to study both intracellular phosphorylation patterns of important signal transduction molecules concomitantly with T cell surface marker expression. A multi-parametric analysis may allow the quantification of the phosphorylation of up to five signaling molecules in CD4 and CD8 T lymphocytes and their naïve, central memory, effector memory, and TEMRA subsets. This enables precise identification of subset-specific signaling and alterations of signaling pathways in physiological and pathological situations. The importance of such detailed analysis is discussed. © 2013 International Society for Advancement of Cytometry

The study of T lymphocyte signaling has been very useful for understanding physiological processes such as cell activation or pathological situations involving T lymphocytes (1, 2). The most common tool for these studies is Western blotting (3). It is a reliable technique but has some drawbacks such as the need for high cell numbers (>2 × 106 cells), difficulties in precise quantification of the signal, the impossibility of distinguishing the status of signaling molecules in cell subsets without sorting or separating beforehand, and its unsuitability for high-throughput analysis. Moreover, sensitivity is an issue when the signal is present in only a small fraction of the cells in the test population. Regarding quantification in cell subsets, recent advances in immunology highlighted the presence of several T cell subsets (actually differentiation stages) often distinguished using the presence of the phosphatase CD45RA and the chemokine receptor CCR7 at the cell surface in combination with other receptors (4, 5). The naïve (CD45RA+CCR7+) and central memory (CD45RA−CCR7+) cells express both CD28 and CD27. The effector memory (CD45RA−CCR7−) and TEMRA (CD45RA+CCR7−) T lymphocytes lose expression of CD27 and CD28; cells with these phenotypes may be present at frequencies as low as 1% (6). This typical example is also true for other cell types such as dendritic cells (7) and may be applied to all leukocytes (8). It is expected that the number of identifiable subpopulations will increase with the advent of recently developed mass spectrometry-based cytometers (9).

Flow-cytometry in general is a state-of-the-art tool to quantify these populations (10). It is also of high efficacy to quantify signal intensities, it requires minimal numbers of cells and it is not as labor-intensive as blotting. Multi-parametric flow cytometry has become possible during the last decade due to technological improvements and also to the increasing availability of natural and synthetic fluorochromes (11). Recently, a range of antibodies binding to phosphorylated moieties of target molecules (phospho-specific antibodies) has become available (12). However, the use of these phospho-antibodies in a multi-parametric setting for combined analysis of cell subsets and simultaneous assessment of signaling activity has been challenging (13). Here, we describe a method developed to quantify the phosphorylation status of one or more signaling molecules in naïve and memory CD4 and CD8 primary lymphocytes in a rapid and reliable manner. The importance of investigating further such signaling events in health and diseases is discussed.

Materials and Methods

Reagents and Cells

Blood samples obtained from healthy middle-aged individuals at the University Hospital blood bank were processed using FicoLite-H (Linaris, Wertheim-Bettingen, Germany) to separate peripheral blood mononuclear cells (PBMC) and cryopreserved. These are healthy donors screened for HIV. The University of Tuebingen's ethical committee approved blood sampling. PBMC were thawed and resuspended (1 to 5 × 106 cells/aml) in serum-free culture medium X-Vivo 15 Lonza (Basel, Switzerland) and extensively washed in PBS containing 2% FCS, 2 mM EDTA, and 0.01% sodium azide (PFEA) which is the working solution for the rest of the experiments. Antibodies used for flow-cytometry experiments such as CD3-FITC, CD3-Alexa-700, CD3-PercP, CD4-PercP, CD4-APC, CD4-PB, CD8-PercP, CD8-APCH7, CCR7-APC, CCR7-PECy7, CD45RA-APC, CD45RA-PE, CD45RA-PB, CD27-FITC, CD27-APC, CD28-Alexa 700, CD28-APC, CD28-PercPCy5.5 were purchased from BD Biosciences (San Jose, CA). Secondary goat anti-mouse Pacific orange was purchased from Invitrogen/Life Technologies (Grand Island, NY). Antibodies used to detect the phosphorylated form of signalling molecules: p-Lck-Alexa647, p-ZAP70-Alexa488, p-p38-PB, p-Tyr-PE, p-ERK1/2-PECy7 as well as buffers for fixation (Cytofix/Cytoperm®) and permeabilization (PermIII®) were obtained from BD Bioscience. Bis-sulfosuccinimidyl suberate (BS3) was obtained from Sigma-Aldrich (St. Louis, MO).

Cell Stimulation

PBMC were stimulated with H2O2 (5 mM) for 15 min at 37°C or with anti-CD3 and anti-CD28 antibodies (5 μg/ml) on ice for 15 min, washed, then crosslinked with a goat anti-mouse Ig (5 μg/ml) for 15 min on ice. After washing out the unbound antibodies, PBMC were transferred to 37°C for 1–15 min. The reaction was stopped by transfer to 4°C and quick spinning of the samples. Non-stimulated PBMC followed the same procedure (washing steps and centrifugations) to have similar conditions. During the activation or resting step, cells were concomitantly incubated with 1 μl Redvid (Invitrogen), a live/dead marker, as per the manufacturer's recommendation. Some experiments were prepared in duplicate: one was placed in the Hypoxia Workstation Invivo2 from Ruskinn Technology (IUL Instruments GmbH, Königswinter, Germany), in which pO2 was controlled by continuous injection of an appropriate amount of N2 to reach 2% oxygen supplemented with 5% CO2. The other was placed in a standard incubator, where cells grow in air with no addition of N2 but supplemented with 5% CO2. The oxygen concentration in the incubator was controlled continuously, and all reagents and plastic materials required were stored inside the hypoxia chamber to minimize contamination with air.

Staining Procedure

Isolated PBMC (1 × 106) were washed twice with PFEA and stained in 50 μl total volume with single or antibody master mix. After 20-min incubation on ice, cells were washed twice with PFEA prior to acquisition. Crosslinking of protein–antibody complexes was performed by mixing an equal volume of freshly prepared stock solution of BS3 (2 mM in PBS) and cells. After 30 min on ice, the reaction was stopped using 100 mM Tris-HCl, pH 7.0, 150 mM NaCl, and washed subsequently with PFEA. Further incubation with Fix/PermIII buffer was applied for some samples.

The working phosflow staining procedure: Resting or activated PBMC were centrifuged and supernatant discarded. Cells were fixed using 250 μl Cytofix/Cytoperm (BD Biosciences) for 15 min on ice. This buffer is usually recommended for intracellular cytokine staining but we found it less harmful than other buffers for the cells and the subsequent steps. Cells were washed twice with Perm/Wash and stained for some surface markers. Cells were washed with staining buffer after 30 min incubation at 4°C. Cell permeabilization was performed on ice for 30 min using the PermIII buffer (BD Biosciences). Cells were then again washed twice with staining buffer and centrifuged to discard supernatant. Finally, cells were stained for phosphorylated intracellular signalling molecules in combination with some surface markers that did not tolerate the permeabilization step. After 30 min at room temperature in the dark, cells were washed with staining buffer and collected for analysis.

Data Processing

For each experiment, mouse or hamster/rat κ-chain CompBeads were stained with the same fluorochrome-labeled antibody used for cell staining and incubated for 20 min at 4°C in the dark. Negative CompBeads were used as unstained negative controls. After washing with PFEA, the cells or beads were resuspended in 200 μl PFEA and measured using an LSR-II flow cytometer equipped with three lasers (405, 488, and 635 nm) and a set of filters to measure up to 11 fluorochromes. The spectral overlap between all channels was calculated automatically by the BD FACSDiva software, after measuring negative and single-colour controls. Data were processed and analyzed using FACSDiva and Flowjo 7.2.2 (Treestar Inc., Ashland, OR) following recently updated guidelines. The data analysis consisted of the sequential gating of the target populations. The SSC-A vs. time plot allowed us to find any change in flow rate while the FSC-A vs. SSC-A plot allowed us to gate the lymphocytic populations. Further exclusion of cells was performed using the CD3 vs. Redvid plot which allowed us to select only viable CD3+ T cells, and FSC-A/FSC-H was used to exclude doublets. T cell subset analysis including frequencies and mean fluorescence intensities was based on this first gating strategy. Fluorescence minus one controls were previously set up to define positive vs. negative populations.

Pitfalls for T Cell Signaling Investigations by Cytometry

Our initial studies were aimed at defining the phosphorylation status of several molecules in resting and activated cell subsets (14). However, this is not possible on a small scale using the most widely applied technique, i.e. Western blotting, because this requires too much material and cell sorting prior to stimulation and analysis. A very good alternative is the use of flow cytometry which requires fewer cells and allows the quantification of the signal (via the mean fluorescence intensity). Our aim was to measure phosphorylation of signaling molecules in T cell subsets using directly conjugated antibodies. However, the manufacturers' protocols (from most companies) only report on single antibody staining of cell lines (usually Jurkat cells as a model for T cells) or PBMC. When applying the provided protocols it has been difficult to successfully use the commercialized reagents and buffer. In Figure 1A we provide representative examples of fluorescence intensities of various surface markers following staining when using the permeabilization buffer “PermIII” from BD Biosciences (in gray) compared to the same staining without permeabilization (in black). The exact composition of the buffer is not in the public domain, but largely relies on alcoholic compounds while other buffers for permeabilization, for example, include saponin. We observed that only some surface staining retained a satisfactory level of fluorescence intensity when using most of commercially available reagents. This was the case when using CD4-Pacific blue (Biolegend, San Diego, CA) or CD8-APCH7 (BD Biosciences) and to a lesser extent CD45-APC (BD Biosciences). However, some surface markers showed a mild to severe reduction in intensity, namely, CD3-PercP (BD Biosciences), CCR7-PECy7 (BD Biosciences), and to a lesser extent CD27-FITC (BD Biosciences). Antibodies directed against various surface markers with different fluorescent dyes and from different suppliers were tested and revealed that depending on the antibody clone and/or the associated-dye, the results greatly vary. Unfortunately, this has to be tested for each antibody as discrepancies exists between expected data and the reality. For this study, we selected reagents from BD Biosciences because of the greater range of dye-conjugated phospho-antibodies available from this manufacturer.2

Figure 1.

Impact of permeabilization on intensity of various surface markers. A: Untreated samples (black) versus PermIII treated samples (gray) are shown. Cells were stained with directly conjugated antibodies: CD4-Pacific blue (PB), CD8-APCH7, CD45-APC, CD3-PercP, CCR7-PECy7, CD27-FITC. A representative histogram is displayed showing the different effects of permeabilization. B: The crosslinking with BS3 did not permit rescue of the signal of Fix/PermIII-treated cells. [Color figure can be viewed in the online issue which is available at wileyonlinelibrary.com]

Figure 2.

Protocol established to measure phospho-signaling molecules in T cell subsets.

Nonetheless, these experiments indicate that it was not possible to standardize the measurement of phosphorylation in T cell subpopulations using some of the surface markers we tested. We tried to identify the reasons for such discrepancies. For this purpose, we used an antibody crosslinker, BS3, and tested whether cell treatment with PermIII buffer induced antibody-epitope dissociation (15). In Figure 1B we show the regular staining for CD3 (control) and the same staining but including fixation and permeabilization in the procedure (Fix/PermIII). The loss of the signal is clear. When adding the crosslinker alone following regular antibody staining, we found that this has no effect (BS3), as expected. However, the crosslinker was not able to maintain the fluorescent signal following fixation and permeabilization (BS3 + Fix/PermIII). This clearly suggests that the procedure alters some of the fluorescent conjugates or that the procedure is too aggressive.

Detection of T Cell Signaling

Since our aim was to standardize the method to measure signaling in T cell subsets, and in cellular subpopulations in general, we set up a protocol testing a wide range of antibodies in this procedure and investigated several possibilities such as surface staining at different steps during the procedure. This allowed us to determine that some antibodies and/or dyes were very sensitive to the procedure, which explains the lack of consistency. To avoid this, we have identified these and now provide the following protocol (Fig. 3) including the step where the corresponding antibody (all from BD Biosciences) can be added to successfully obtain a similar signal compared to a regular staining. Note that not only directly-conjugated antibodies but also indirect staining were tested (mouse anti-human CCR7 and goat anti-mouse Pacific Orange from Invitrogen). The proposed protocol requires 1 × 106 PBMC and allowed us to clearly distinguish CD3+ T cells (Fig. 3A, lower left). Resting, CD3/CD28-activated, and H2O2-treated PBMC were stained according to our protocol using a single surface marker (CD3); we aimed to measure the phosphorylation status of up to four different signaling molecules simultaneously (Lck, ZAP-70, ERK1/2, and p38) along with the overall tyrosine phosphorylation status (16). The data presented suggest that it is possible to combine up to five different phospho-specific antibodies in a single experiment (Fig. 3A). Dead cells were gated out using the exclusion dye RedVid to ensure that the signal did not derive from non-specific staining (17). The CD3/CD28-activated (gray histograms) and H2O2-treated T cells (black histograms) show a positive signal for Lck, ZAP-70, ERK1/2, and p38 compared to resting cells (white histograms). There was a differential activation of the signaling molecules according to the stimuli used (CD3/CD28 or H2O2) and depending on the targeted molecule. Thus, there was a very high intensity for the overall tyrosine phosphorylation of T cells treated with H2O2 compared to CD3/CD28 (Fig. 3B). Also, the phosphorylation of Lck was more induced than that of ERK-1/2 in H2O2-treated cells. Overall, this suggests that detection of multi-phospho-signaling is possible for a specific population. The assay is robust as indicated in Figure 3B. Three different staff used the same protocol in order to stain CD4+ T cells from the same sample using the PermIII buffer. The frequency of CD4+ T cells within the CD3+ population was very similar for the three experiments (44.1, 42.2, and 45.9%, left panel). PBMC were stimulated with anti-CD3 and anti-CD28 mAbs followed by crosslinking as described, split in two after fixation and stained in two different experiments by the same staff, and measured the same day (right panel). The histogram shows a similar phosphorylation profile of ERK1/2 (solid line) in two independent experiments (p-ERK1/2 MFI: 691 and 729). The robustness and standardization of the measure is very important as phosphorylation is dependent on intensity of stimulation, time, and is often represented by a shift in mean fluorescence intensity (Fig. 3A).

Figure 3.

Multi-phospho-signaling measurement in T cells. A: H2O2 and CD3/CD28 activated T cells were stained for CD3 and phospho-Lck, -ZAP70, -p38, -Tyr, and -ERK1/2. B: The same sample was processed by three different individuals to assess CD4 expression using our protocol. Samples were acquired the same day using the same cytometer (left). The same sample (activated T cells) was stained by the same staff on two different occasions and acquired the same day on the same cytometer. Phosphorylation of ERK1/2 is shown for CD4+ T cells.

Polychromatic Flow Signaling

After validating the protocol, we explored the possibilities for investigating signaling in cell subpopulations. Most published T cell studies now include distinctions between behaviors of the different subsets. A common way to achieve this is to use the RA isoform of the phosphatase CD45 and the chemokine receptor CCR7 to identify four major populations in double staining (5). Thus, we stimulated PBMC with anti-CD3/CD28 mAbs and stained them for CD3, CD8, CCR7, and CD45RA and with an antibody specific for the phospho-tyrosine residues, enabling us to identify the overall tyrosine phosphorylation and thus activation status of the cell. As shown in Figure 4, the naïve, central memory, effector memory, and TEMRA subpopulations were easily identified within the CD3+CD8+ subset (top panel). The overall phosphotyrosine pattern of resting cells was similar in all four subsets but was different in the activated cells. We found the highest phosphorylation level in the naïve and central memory subsets (Fig. 4A), i.e. those expressing the highest levels of the co-stimulatory molecule CD28 (data not shown). In more differentiated cells such as effector memory and even more so in TEMRA cells, the phosphorylation level is lower. This is most likely due to the increased frequency of CD28-negative cells in these two subsets (18). While 95–100% of naive and central memory cells express CD28, effector memory cells lacking its expression may vary from 15% to 45% in different individuals. This is even more pronounced in the case of TEMRA cells which are considered to represent a later stage of T memory cell differentiation. For this latter subset, up to 80% of cells may lack CD28 expression. We suggest that lower phospho-signal detected is due to lower CD28 expression impacting on TCR/CD28 signaling crosstalk necessary for T cell activation. Still, a portion of the TEMRA population is able to up-regulate phospho-signaling. Although this particular pathway is altered, the use of non-specific stimulation such as PMA/Ionomycin would enable the receptor to be bypassed and could identify the overall capacity of the cells, independently of receptor expression. The results indicate that the phosphorylation levels following activation can be subset-specific and that measured differences in activation levels between individuals or groups of individuals may simply reflect differences in distribution of the subpopulations and their different levels of surface markers, some of which are important for the regulation of signal transduction.

Figure 4.

Phospho-signaling in multiple subsets and conditions. A: PBMC were stimulated with anti-CD3/CD28 and stained for the classical naïve/memory subset distribution and p-Tyr in CD8+ T cells. Resting and activated cells are shown with percentage of p-Tyr+ T cells in each subset. B: Resting (top) and activated (middle) CD4+ T cells were stained for ERK1/2, Lck, and Tyr phosphorylation. The p-Tyr profile of p-Lck+ CD4+ T cells with/without p-ERK1/2 is also shown (bottom). C: The low T cell metabolism in low tension (2% oxygen) correlates with low p-Tyr profile in CD3+ T cells compared to atmospheric oxygen (20%).

In another set of experiments, we tested T cell responses to stimulation and performed a multi-parametric analysis in order to demonstrate the feasibility and value of such investigations. In Figure 4B, we show an example of how such an analysis can be performed. A dot plot for ERK1/2 and Lck phosphorylation status of gated CD3+CD4+ T cells is shown in resting (upper panel) and activated cells (middle panel). The p-ERK1/2+ p-Lck− and p-ERK1/2+ p-Lck+ phosphorylated cells were then transposed to a histogram for their overall phosphotyrosine levels. The p-ERK1/2+ p-Lck+ cells are those displaying the highest p-Tyr levels, suggesting that ERK1/2 phosphorylation is an important driver of overall T cell activation status. This also shows that correlations between signaling events can be made. We have previously reported (19) that T cell functions measured in vitro are crucially influenced by the oxygen level to which they are exposed, and we show here (Fig. 4B), that the reduced proliferative and differentiation response of CD3+ T cells observed was in part due to reduced phosphorylation following their activation at low oxygen (2% O2). While the mean p-Tyr intensity of CD3+ T cells following 2 days stimulation at 20% O2 was 3,811, this value was 4.7-fold lower in CD3+ T cells cultured at 2% O2. Again, our protocol shows consistency in the determination of phosphorylation pattern and data presented correlate with published data. In conclusion, the study of T cell signaling using this protocol allows the determination of subset-specific multi-phosphorylation profiles.

T Cell Signaling in Health of Disease: The Past and Future Contribution of Flow Cytometry

We have demonstrated that using this simplified protocol we could measure T cell signaling events in different cell populations in a reproducible manner. The combination of surface markers with intracellular markers has already been shown for cytokines (20) but here we provide evidence that combining both protocols allows detection of up to five different signaling molecules in a single experiment at the single cell level. Moreover, four major T cell subpopulations (naïve, central memory (CM), effector memory (EM), and T effector memory re-Expressing CD45RA (TEMRA)) were clearly identified according to the CCR7, CD45RA dual straining model, with no significant change in their distribution using our method compared to standard staining. This will be of value for better dissecting out the role of signaling in certain diseases or the effect of diseases on signaling events (21). A recent study used flow cytometry staining to show that in the physiological steady state naïve and memory T cells display different phosphorylation levels (22). CD4+CD45RA+CD27− memory T cells display higher levels of both the total and phosphorylated p38 protein. This may reveal important information on T cell biology and which pathway may be involved in the maintenance, survival, or senescence of T cells. However, that study was limited to one signaling molecule; increasing their numbers in the assays would allow identification of the earlier events associated with functionality, because cells with high levels of p-p38 in the steady-state display reduced functionality.

We believe that our simplified method (compared to other methods which need many more steps and reagents such as formaldehyde, Triton X-100 or MeOH) using commercially available reagents is an improvement on earlier protocols. The establishment of such protocols aims at reducing the amount of sample and reagents needed; at the moment, it requires as few as 1 × 106 cells per assay. Minimizing the method further was possible (0.5 × 106 PBMC) but within limits, because some T cell subsets are present at low abundance. Thus, detailed T cell signaling investigation is recommended using “untouched/unsorted” cells to avoid unwanted cell activation due to antibody binding (necessary for cell population gating/sorting or enrichment), cell sorting, and cell separation. It is highly relevant for sensitive cells such as polymorphonuclear neutrophils. A Pubmed search reveals signaling studies involved more than 10 times more T cells than neutrophils. In addition to the scientific interest, a logistical advantage is in overcoming the difficulty of working with neutrophils due to their susceptibility to activation. Basic experimental procedures such as Ficoll-density centrifugation may already activate these cells and hence alter the signaling events one would investigate. Thus, further processing of these cells, such as by cell sorting, would interfere with their steady-state phosphorylation status even more. Our protocol is one amongst several, and developments and improvements will certainly continue to be proposed. We have previously shown that the use of other protocols also leads to acceptable data (23). More urgently, a consensus in the protocols is necessary as different protocols may lead to discrepancies. Testing, standardization, and validation is welcomed, as it has been fairly successful for other measurements (24).

On the whole, analysis of phosphorylation events by flow cytometry allows studies to be carried out that would otherwise fail to identify whether T cell signaling is altered overall or at the cell subset level. A study showed that age-associated changes in heat shock protein (Hsp) expression in leukocytes is not similar in all subsets (25). The basal Hsp70 levels in peripheral mononuclear cells were analyzed by flow cytometry in combination with surface marker expression to discriminate naive and memory cells. The memory cell subsets expressed more Hsp70 than the naïve subset, suggesting that age-related changes in basal Hsp70 levels in leukocytes are linked to the altered frequency of lymphocyte subsets and not to increases in Hsp70 per se. This is a typical example of the need to investigate phenomena such as T cell signaling at the subset level. A major limitation of this present method remains the lack of information on the localization of the phosphorylated signaling molecules. However, technologies combining cytometry and microscopy allowing the localization of fluorescent signals exist (26). Improvements on the technical side will be of major importance and will enable to us answer major questions such as which signaling pathway is involved in regulatory T cell function? What drives Th17 differentiation? Is there differential signaling leading to activation of transitional B cells? Which signaling pathway is involved in the survival of neutrophils? Which signaling event is promoting senescence? anergy? Doody et al. have reviewed the role of protein tyrosine phosphatases (PTP) in T cell physiology and implication in diseases (27). They have determined that PTP are involved in systemic inflammation, cytokine signaling, hematopoiesis (myeloid, T cells, B cells, stem cells). Moreover, a PTPN2 polymorphism (rs2542151) was identified which is associated with Type 1 diabetes, rheumatoid arthritis, and Crohn's disease (28). The correlation of this SNP with the co-occurrence of all three diseases was significant which suggests that PTP is involved in common pathogenic pathway. Associations with other diseases such as ulcerative colitis suggest the use of PTPs as putative targets for anti-inflammatory drugs. However, as noted by Doody et al., it is not clear whether PTP expression is modified in such pathologies and if their activation status (assessed by phosphorylation) and activity is playing a role or is merely influenced by the disease. We have also identified age-associated PTP changes. The Src homology domain-containing protein tyrosine phosphatase-1 (SHP-1) shows dysregulated activity and localization in the membrane rafts in neutrophils from elderly individuals (29). Still, whether this is specific for a subset of neutrophils was not clear. With the advent of flow cytometry such questions could be answered. Although the recent introduction of advanced mass spectrometry-based flow cytometry is an advantage for the flow community and scientists, the cost and still-in-development phase of the equipment and reagents is a barrier for the majority of flow cytometry users. Still, fluorescence flow cytometry is the best tool to answer these major questions.

Ancillary